Abstract
Lorem Ipsum….
Rade: TBD.
Switzerland has a long tradition of democracy, with its modern
political system rooted in the 1848 Federal Constitution. The modern
Swiss State, along with its constitution and the bicameral legislative
system has been established in 1848 (Church,
2013, Chapter 6). Swiss elections take place every four years,
with the federal government following a system of direct democracy that
allows citizens not only to elect representatives but also to
participate in referendums and initiatives. As mentioned before,
Switzerland’s Federal Parliament consists of two chambers: the National
Council (200 members) and the Council of States (46 members). The
National Council represents the population proportionally, while the
Council of States represents the cantons, with each full canton electing
two representatives and six half-cantons electing one. This bicameral
system aims at balancing demographic and regional in legislative
decisions.
According to the annual overview of the federal office for statistics
(Statistik, 2023), Switzerland has a
population of 8.8 million people (p. 132), of which around 74% hold
Swiss citizenship. The remaining 26% are non-Swiss residents, including
permanent residents, cross-border workers, and asylum seekers (p. 142).
Swiss nationality is acquired for most citizens through descent and for
migrants by naturalization, which is known to be a decentralised process
mostly relying on cantonal and municipal approval. Only a fraction of
foreign nationals applies for naturalisation. In the last decade,
between 30’000 to 45’000 people a year, which corresponds to around 2%
of the population eligible for naturalisation (p. 142).
Non-Swiss residents, despite their significant share of the population,
have limited political rights at the federal level. Some cantons and
municipalities allow foreigners to vote in local elections or even run
for office, but they are largely excluded from national decision-making
(jura.ch, 2025; ne.ch, 2025). Politics of
Switzerland are described as relatively polarized compared to
international standards (Jansen & Stutzer,
2024, p. 3)
While many European countries have seen an extraordinary rise of
right-wing politics throughout the 2010s, not much had shifted in
Switzerland. The right-wing SVP (Schweizerische Volkspartei, in french:
Union démocratique du centre UDC) has been part of the governing
coalition since the 1990s. Certain segments of the party have been
classified as far-right and right-wing extremist, most notably due to
the party’s hardline positions on migration and strict opposition to any
legal protection of social, cultural and religious minorities like swiss
muslims or LGBT citizens (Ellermann, 2021, p. 3
& p.102; Jansesberger & Rhein, 2024, pp. 3–5).
The relationship between regional demographic composition and voting
behavior has been a subject of ongoing research and debate, particularly
if areas with lower proportions of migrants tend to support stricter
immigration policies compared to ethnoculturally diverse regions. In
this study, we examine the most recent Swiss election to analyze
potential correlations between the percentage of non-Swiss residents and
the electoral performance of the Swiss People’s Party (SVP), which
centered its campaign on an anti-immigration platform. To control for
cofounding variables, other demographic markers as age and education
will be included.
All datasets used in this analysis were obtained from the Swiss Federal Office of Statistics (BFS). Each dataset included the BFS municipality ID, with the exception of the dataset on education levels by district, which lacked an ID. Matching this dataset by district name using regular expressions was not feasible due to inconsistencies in district naming conventions. To address this, district numbers were manually added to the table. Aside from this exception, no further modifications were made to the datasets.
| Dataset Description | Dataset ID |
|---|---|
| Election Results 2023 | sd-t-17.02-NRW2023-parteien-appendix.csv |
| Citizenship Pecentage | px-x-0102010000_104_20250127-155044.xlsx |
| Education | su-e-40.02.15.08.05-2022.xlsx |
| Citizenship acquisition | px-x-0102020000_201_20250129-134648.xlsx |
| Age distribution | su-d-01.02.03.06.xlsx |
| Some income or wealth metric TBD | TBD |
| Datatable Communes, Districts and Cantons | Gemeindestand.xlsx |
Over the past several decades, the distribution of dominant political parties in Swiss national elections has remained relatively stable. The primary parties represented in the federal parliament, along with their respective campaign focuses in the most recent elections, are detailed below (bpb.de, 2023) . For improved readability, this paper will rely on german-language party names.
| Party | 2023 Campaign Focus |
|---|---|
| SVP | Right-wing, anti-immigration, anti-welfare, free market policies |
| SP | Left-wing, pro-welfare, pro-worker policies, reducing cost of living |
| FDP | Center-right, free market policies and improved access to international markets |
| Die Mitte | Conservative centrist, pro-defense, tax cuts for married couples |
| GLP | Progressive centrist, climate protection, EU alignment, liberal market policies |
| GPS | Left-wing, climate protection, biodiversity, state regulation of business |
Regional parties like the far-right Lega in the Canton of Ticino or Movement Citoyen Romande in Geneva have been taken into account for some analysis. Due to their regionality the data for them is scarse and might not yield very reliable results.
The most recent parliamentary elections in Switzerland took place on
October 22, 2023. Switzerland’s bicameral parliament comprises the
National Council, which proportionally represents the Swiss population
with seats allocated to each canton based on its population, and the
Council of States, where all cantons have equal representation. As the
National Council reflects the population proportionally, its election
results are often regarded as an indicator of trends in public opinion
regarding politics and policy.
As per Januar first 2024 Switzerland has 2XXX registered municipalities, that were taken into account for this analysis.
During prepocessing of data there were minor inconsistencies since the used datasets span from 2019 to 2024 and some smaller municipalities had merged with larger communes during those years. Therefore, some smaller communities can not be represented perfectly in the analysis. In 2020, the former municipalities of Vogorno, Sonogno, Corippo, Brione (Verzasca) and Frasco merged to form the new municipality of Verzasca. In 2021 many small communes disappeared from the map: Haldenstein (merged into Chur), Bauen (merged into Seedorf), Surpierre (merged into Cheiry), Charrat (merged into Martigny), Montherod (merged into Aubonne), Casti-Wergenstein, Donat, Lohn and Mathon (merged together),Miège, Venthône and Veyras (merged together), municipalities of Gänsbrunnen and Welschenrohr (mmerged together),Obersteckholz (merged into Langenthal), Gettnau (merged into Willisau), Altwis (merged into Hitzkirch), Rohr (merged into Stüsslingen), Bagnes and Vollèges (merged together), Corcelles-Cormondrèche, Peseux and Valangin (merged into Neuchâtel), Les Brenets (merged into Le Locle), Bioley-Orjulaz (merged into Assens), Apples, Cottens, Pampigny, Sévery, Bussy-Chardonney and Reverolle merged into the new municipality of Hautemorges. In 2022, the following communes were dissolved: Galmiz, Gempenach and Clavaleyres (merged into Murten), Essertes (merged into Oron), Bözen, Effingen, Elfingen and Hornussen (merged together), Bad Zurzach, Baldingen, Böbikon, Kaiserstuhl, Rekingen, Rietheim, Rümikon and Wislikofen (merged into the municipality of Zurzach), Rüte and Schwende (merged together), Blonay and Saint-Légier-La Chiésaz merged together. 2023 saw the dissolution of Herznach and Ueken (merged togehter), La Chaux-des-Breuleux (merged into Le Breleux), Burg (merged into Menziken), Adlikon and Humlikon (merged to form the municipality of Andelfingen (merged together), Hemberg and Oberhelfenschwil merged into the new municipality of Neckertal. In 2021, Lüterswil-Gächliwil merged into Buchegg, Turgi was merged into Baden. Bonfol merged with Beurnevésin, Wangenried merged into Wangen an der Aare.
Even tough the list of communes seems long, most of these munnicipalities do not affect the outcome of the weighted correlation since their population is in almost all cases below 1000 and in many cases below 100, with Corippo TI having a population of to 9 inhabitants at the time of the merger being the lowest populated commune ever recorded in Switzerland. Therefore these mergers do not affect the outcome of the analysis in any significant way.
All the data points (apart from education level) were available on municipality-level. To calculate the corresponding values for each canton, the municipal values were grouped by canton and weighted by population size of the municipality in case of the demographic values or by vote numbers in the case of the election results. In the same way a total for all of Switzerland was calculateds and included in some plots to better compare the cantons not only in relation to one another but to the swiss average values.
For our analysis we have focused on the following five demographic indicators below (s. Table). The five indicators are of particular interest for measuring their influence on far-right politics, since previous research has shown them to be a predictor of far-right party’s success.
| Indicator | Descrption | Metric |
|---|---|---|
| Non-Swiss Population | People residing in Switzerland without holding Swiss Citizenship | % of total population |
| Citizenship Aquisition | Rate of Migrants acquiring Swiss citizenship in 2023 | % of migrant population |
| Income per Capita | Normalized taxable income. | in Swiss Francs |
| Education Level | Highest educational achievement. | 3 Grades (Low, Secondary, Tertiary) |
| Age Ratio | Percentage of people aged >64 in relation to population aged 18-64 | % of working age population |
The indicator “Non-Swiss Population” represents the percentage of citizens holding another passport than the swiss one per XX.XX.DATE in relation to the base number of the entire population. Migrants that have acquired swiss citizenship count as swiss population are not disclosed or analysed separately.
The indicator “Citizenship Aquisition” represents the percentage of citizens acquiring swiss passport in 202X in relation to the base number of the entire migrant population. The rate of absorption of foreign nationals into the swiss society indicates both willingness to become Swiss citizen from a migrant perspective as well as the willingness of local communities to support non-swiss population’s aim to integrate on a citizenship level.
The Income per capita is a normalized measure used by the Bundesamt für Statistik to compare income across Switzerland. It represents the income (minus deductibles on federal level) that is recorded by the federal tax office in collaboration with the cantonal tax administrations. Therefore it does not represent the net income, but rather the taxable income.
Education levels, usually measured in form of the highest achieved school degree, are crucial for economical and social development of populations. They significantly impact individuals’ perception of life and political leanings. The three indicators are as follows: - Low Education: Only obligatory school education, which in Switzerland represents usually 9 school years. - Secondary Education: A professional college degree, usually acquired as a result of an apprenticeship (Berufslehre, Apprentissage) - Tertiary Education: An academic university degree, acquired as a result of a university study program but also includes higher degrees from professional colleges (Höhere Fachschule, Ecole Superieure)
The age ratio is a standardized measure of the Bundesamt für Statistik, that measures the percentage of aged populations (65 and above) in relation to the adult respectively working-age population (18-64). Therefore, high number indicates an outsized aged population.
Initially the correlation between migrant population and far-right election success was calculated both weighted and unweighted, meaning each municipality counts as one unit nevertheless of the population size. The unajusted analysis did not produce a significiant correlation, presumebly, because both left-wing votership and migrant populations tend to be concentrated in larger cities and therefore could not be represented in an unweighted statistical design. But when adjusted for population size, there were clear correlations visible.
The statistical analysis relies on correlation and linear models. In orner to improve the model’s performance, the demographic indicators were scaled and weighted. The scaling improves interpretability, numerical stability and measureability of smaller differences while eradicating the influence of outliers. The weighting was necessary for practical reasons since the size of a municipality also can be an predictor of election outcomes and if every municpality couunted as one unit this would suppress the outcomes of bigger citites.
Most of the plots did not include statistical models but do rather visalise the election outcome unmodelled.
Explanatory Analysis
In order to better understand the results of the analysis we assessed the data set. The main goal was to get an idea of the existing trends and distributions to reason why certain effects might cause correlations or regressions.
The graph visualises the proportion of non-Swiss citizens per canton as well as all of Switzerland. This information is of importance since previous studies (Sources) show that areas, which do not have large populations of migrants, tend to vote for parties that run on anti-migration campaigns. This highlights an important contradiction, meaning that people who are not exposed to migrant populations tend to be more critical towards migrations, than votes in areas with higher ethno-cultural diversity.
The graph visualises the proportion of different educational layers in the demography. We know from other studies (Source here) that most Swiss people do have in fact a Secondary Education degree, which leads us to the assumption that the demographic segment only in possession of an obligatory education might contain a high percentage of non-Swiss populations, particularly refugees which could not complete their schooling programme since they have been forced to leave by war and conflict. Therefore, the Swiss population can be divided roughly in secondary and tertiary educated people. The populations having a secondary-level education are of particular interest since they represent mostly Swiss and Swiss-born migrants, with or without Swiss citizenship, which did not further continue education after receiving their professional degree. From prior studies on this topic, we know that they might be more likely to vote in favour of the far-right party SVP and less likely to vote in favour of left-wing SP and Greens as well as centrist parties GLP and FDP. From the graph we can extract the information that we see some consistency in educational level across Switzerland with some fluctuations from XX on the lower end up to XX on the higher end.
Plots TBD
…
The plot shows the age ratio of the municpalities and the election outcome of SVP in 2023.
The plot shows the taxable income of the municpalities and the election outcome of SVP in 2023.
The plot shows the percentage of non-swiss population of the municpalities and the election outcome of SVP in 2023.
The plot shows the naturalization rate in percent of the municpalities and the election outcome of SVP in 2023.
The plot shows the population size in number of inhabitants of the municpalities and the election outcome of SVP in 2023.
The plot shows the education level in percentage of inhabitants only holding a secondary eductation degree of the municpalities and the election outcome of SVP in 2023.
Maps with municipalities (or cantons, districts etc.).
Procedural limitations of the analysis include the unreliability of election results to establish a causal relationship to voting motivation. Since the far-right SVP campaigned on an anti-migration platform we assume that many voters were mobilised by this issue. But this does not account for non-voters, which represented in 2023 XX% of the population. Also, some people might have been voting for far-right parties for other reasons like fiscal austerity and the far right’s opposition to social liberty policies - sometimes called “culture war” issues. And as always in elections, a lot of people stick for many years with the party they most identify with. The election data doesn’t include hints to non-voters motivation to abstain from voting and therefore may not be representative of the entire population, leading to biased results.
Some limitations must be noted for the statistical analysis. For the education levels BfS only publishes data per district, not per municipality. The scarcity of data points means a smaller reliability of the indicator. Demographic factors like age ratio, income per capita, and migrant population percentage can be highly correlated, making it difficult to isolate the effect of each variable and leading to multicollinearity. The relationship between election results and demographic factors may not be linear. Linear models may oversimplify these relationships. Also, the chosen demographic indicator might not be conclusive and failing to include relevant variables can lead to biased estimates. The variance of errors may not be constant across observations, violating one of the key assumptions of linear regression (Heteroscedasticity).
Details of the analysis.
How to interpret the results.